The effect of reinforcement probability on time discrimination in the midsession reversal task

Giving time a chance in the midsession reversal task

The midsession reversal task involves a simultaneous discrimination between stimuli S1 and S2. Choice of S1 but not S2 is reinforced during the first 40 trials, and choice of S2 but not S1 is reinforced during the last 40 trials. Trials are separated by a constant intertrial interval (ITI). Pigeons learn the task seemingly by timing the moment of the reversal trial. Hence, most of their errors occur around trial 40 (S2 choices before trial 41 and S1 choices after trial 40). It has been found that when the ITI is doubled on a test session, the reversal trial is halved, a result consistent with timing. However, inconsistent with timing, halving the ITI on a test session did not double the reversal trial. The asymmetry of ITI effects could be due to the intrusion of novel cues during testing, cues that preempt the timing cue. To test this hypothesis, we ran two types of tests after the regular training in the midsession reversal task, one with S1 and S2 choices always reinforced, and another with S1 always reinforced but S2 reinforced only after 20 trials when the ITI doubled or 40 trials when the ITI halved. For most pigeons, performance was consistent with timing both when the ITI doubled and when it was halved, but some pigeons appeared to follow strategies based on counting or on reinforcement contingencies.

Midsession reversal task with starlings: A quantitative test of the timing hypothesis

In the Mid-Session Reversal task (MSR), an animal chooses between two options, S1 and S2. Rewards follow S1 but not S2 from trials 1-40, and S2 but not S1 from trials 41-80. With pigeons, the psychometric function relating S1 choice proportion to trial number starts close to 1 and ends close to 0, with indifference (PSE) close to trial 40. Surprisingly, pigeons make anticipatory errors, choosing S2 before trial 41, and perseverative errors, choosing S1 after trial 40. These errors suggest that they use time into the session as the preference reversal cue. We tested this timing hypothesis with 10 Spotless starlings. After learning the MSR task with a T-s Inter-Trial Interval (ITI), they were exposed to either 2 T or T/2 ITIs during testing. Doubling the ITI should shift the psychometric function to the left and halve its PSE, whereas halving the ITI should shift the function to the right and double its PSE. When the starlings received one pellet per reward, the ITI manipulation was effective: The psychometric functions shifted in the direction and by the amount predicted by the timing hypothesis. However, non-temporal cues also influenced choice.

Assessing human performance during contingency changes and extinction tests in reversal-learning tasks

Serial reversal-learning procedures are simple preparations that allow for a better understanding of how animals learn about environmental changes, including flexibly shifting responding to adapt to changing reinforcement contingencies. The present study examined serial reversal learning with humans by arranging both midsession and variable contingency reversals across two experiments. We also examined the effects of extinction by adding nonreinforced trials at the end of later sessions and provided the first evaluation of effects of win-stay/lose-shift versus counting strategies on accuracy and response latency of humans' reversal-learning performance. In each experiment, responding tracked contingency reversals, primarily with participants using either win-stay/lose-shift or counting strategies. Introducing variable reversal points in the second experiment resulted in near-exclusive win-stay/lose-shift responding among participants and eliminated counting of trials. Each experiment also revealed an immediate shift from S2 to S1 after experiencing extinction during the initial test trial, indicating resurgence of the initial response through a win-stay/lose-shift response pattern. Therefore, the present study replicates and extends prior findings of a win-stay/lose shift response pattern in situations of greater uncertainty. These findings suggest that differences in environmental certainty induce qualitatively different decision-making strategies.

Step changes in the intertrial interval in the midsession reversal task: Predicting pigeons' performance with the learning-to-time model

Our goal was to assess the role of timing in pigeons' performance in the midsession reversal task. In discrete-trial sessions, pigeons learned to discriminate between 2 stimuli, S1 and S2. Choices of S1 were reinforced only in the first half of the session and choices of S2 were reinforced only in the second half. Typically, pigeons choose S2 before the contingency reverses (anticipatory errors) and S1 after (perseverative errors), suggesting that they time the interval from the beginning of the session to the contingency reversal. To test this hypothesis, we exposed pigeons to a midsession reversal task and, depending on the group, either increased or decreased the ITI duration. We then contrasted the pigeons' performance with the predictions of the Learning-to-Time (LeT) model: In both conditions, preference was expected to reverse at the same time as in the previous sessions. When the ITI was doubled, pigeons' preference reversal occurred at half the trial number but at the same time as in the previous sessions. When the ITI was halved, pigeons' preference reversal occurred at a later trial but at an earlier time than in the previous sessions. Hence, pigeons' performance was only partially consistent with the predictions of LeT, suggesting that besides timing, other sources of control, such as the outcome of previous trials, seem to influence choice.

Pigeons' midsession reversal: Greater magnitude of reinforcement on the first half of the session leads to improved accuracy

In the midsession reversal task, pigeons are trained on a simultaneous two-alternative discrimination in which S1 is correct for the first half of the session and S2 is correct for the second half of the session. Optimally, pigeons should choose S1 until it stops being correct and choose S2 afterward. Instead, pigeons anticipate S2 too early and continue choosing S1 even after the reversal. Research suggests that they attempt to time the reversal rather than use the feedback from the preceding response(s). Recently, there is evidence that performance is almost optimized by generating an asymmetry between S1 and S2. For example, pigeons' accuracy improves if correct S1 responses are reinforced 100% of the time, but correct S2 responses are reinforced only 20% of the time. Similarly, accuracy improves if S1 requires one peck but S2 requires 10 pecks. Accuracy does not improve, however, if the value of S1 is less than the value of S2. In the current experiment, we manipulated the magnitude of reinforcement. For the experimental group, correct responses to S1 were reinforced with five pellets of food and correct responses to S2 were reinforced with one pellet. For the control group, all correct responses were reinforced with three pellets. Consistent with the earlier findings, results indicated that there was a significant reduction in anticipatory errors in the experimental group compared with the control, and there was no significant increase in perseverative errors.

The Midsession Reversal Task with Pigeons Does a Brief Delay Between Choice and Reinforcement Facilitate Reversal Learning?

In a midsession reversal task, the session begins with a simple simultaneous discrimination in which one stimulus (S1) is correct and the other stimulus (S2) is incorrect (S1+/S2-). At the midpoint of the session, the discrimination reverses and S2 becomes the correct choice (S2+/S1-). When choosing optimally, a pigeon should choose S1 until the first trial in which its choice is not reinforced and then it should shift to S2 (win-stay/lose-shift). With this task, pigeons have been shown to respond suboptimally by anticipating the reversal (making anticipatory errors) and continuing to choose S1 after the reversal (making perseverative errors). This suboptimal behavior may result from a pigeon's relative impulsivity due to the immediacy of reinforcement following choice. In other choice tasks, there is evidence that the introduction of a short delay between choice and reinforcement may decrease pigeons' impulsivity. In the present experiment, a delay was introduced between stimulus selection and reinforcement to assess whether it results in a decrease in anticipatory and perseverative errors. Pigeons that had a delay between choice and reinforcement were a bit slower in acquiring the midsession reversal task compared to those without a delay, but showed no decrease in either anticipatory or perseverative errors. It is likely that the pigeons' natural tendency to use time from the start of the session to the reversal as a cue to reverse prevented the delay from increasing accuracy on this task.

The midsession reversal task: A theoretical analysis

In the midsession reversal task, choice of one stimulus (S1) is correct for the first half of each session and choice of the other stimulus (S2) is correct for the last half of each session. Although humans and rats develop very close to what has been called a win-stay/lose-shift response strategy, pigeons do not. Pigeons start choosing S2 before the reversal, making anticipatory errors, and they keep choosing S1 after the reversal, making perseverative errors. Research suggests that the pigeons are timing the reversal from the start of the session. However, making the reversal unpredictable does not discourage the pigeons from timing. Curiously, pigeons' accuracy improves if one decreases the value of the S2 stimulus relative to the S1 stimulus. Another form of asymmetry between S1 and S2 can be found by varying, over trials, the number of S1 or S2 stimuli. Counterintuitively, if the number of S2 stimuli varies, it results in a large increase in anticipatory errors but little increase in perseverative errors. However, if the number of S1 stimuli varies over trials, it results in a large increase in perseverative errors but no increase in anticipatory errors. These last two effects suggest that in the original midsession reversal task, the pigeons are learning to reject S2 during the first half of each session and learning to reject S1 during the last half of each session. These results suggest that reject learning may also play an important role in the learning of simple simultaneous discriminations.

Past outcomes and time flexibly exert joint control over midsession reversal performance in the rat

In a midsession reversal task, subjects choose between two stimuli on every trial; only responses to one stimulus are reinforced. Halfway throughout the session, contingencies are reversed: previously reinforced responses are now extinguished and vice versa. Both, the outcome of the previous trial and the time elapsed since the beginning of the session, may predict the availability of reinforcement and determine choice. Thus, this task has typically been used to study cognitive flexibility and the temporal organization of behavior. This study assessed how past outcomes and time interact for behavioral control when each cue predicts the availability of reinforcement to a different extent. Eight rats were trained in four variations of the midsession reversal task differing in the reliability of outcomes and time as predictors of the reinforced response. We manipulated the reliability of the outcomes by providing either continuous or partial reinforcement, and the reliability of time by fixing the moment of reversal (middle of the session) or making the reversal unpredictable (semi-random trial). Results suggest that behavioral control alternates between outcomes and time according to the relative reliability of each cue. Model simulations show that outcomes and time may jointly determine behavior, and that momentary reinforcement rate may determine their relative influence.

Midsession reversal learning by pigeons: Effect on accuracy of increasing the number of stimuli associated with one of the alternatives

The midsession reversal task involves a simultaneous discrimination in which choice of one stimulus (S1) is correct for the first 40 trials and choice of the other stimulus (S2) is correct for the last 40 trials of each 80-trial session. When pigeons are trained on the midsession reversal task, they appear to use the passage of time from the start of the session as a cue to reverse. As the reversal approaches, they begin to make anticipatory errors, choosing S2 early, and following the reversal they make perseverative errors, continuing to choose S1. Recent research suggests that anticipatory errors can be reduced (while not increasing perseverative errors) by reducing the probability of reinforcement for correct S2 choices from 100% to 20%. A similar effect can be found by increasing the response requirement for choice of S2 from one peck to ten pecks. In the present experiments, we asked if a similar effect could be attained by increasing the number of stimuli that, over trials, could serve as S2. Instead, in both experiments, we found that increasing the number of S2 stimuli actually increased the number of anticipatory errors. Several interpretations of this result are provided, including the possibility that attention to the variable S2 stimuli may have interfered with attention to the S1 stimulus.